The present invention generally relates to task processing.
Traditionally, single-threaded systems are limited by how fast each item can be processed. Further, in processing workflows, with some systems, an error later in the workflow can leave the system in an inconsistent state, or can cause loss of data.
Additionally, scaling traditional systems can be difficult. Physical servers must be shut down to add resources. Virtual machines can scale up (add resources), but traditional systems wouldn't allow scaling out (adding nodes).
A need exists for improvement in task processing. This need, as well as other needs, is addressed by one or more aspects of the present invention.